1 /*
2  * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 #include <sys/types.h>
18 #include <sys/stat.h>
19 #include <sys/queue.h>
20 #include <sys/stdint.h>
21 
22 #include <stdio.h>
23 #include <stdlib.h>
24 #include <string.h>
25 #include <sha1.h>
26 #include <zlib.h>
27 #include <ctype.h>
28 
29 #include "got_error.h"
30 #include "got_object.h"
31 #include "got_commit_graph.h"
32 
33 #include "got_lib_delta.h"
34 #include "got_lib_zbuf.h"
35 #include "got_lib_object.h"
36 #include "got_lib_object_idset.h"
37 
38 struct got_commit_graph_node {
39 	struct got_object_id id;
40 
41 	/*
42 	 * Each graph node corresponds to a commit object.
43 	 * Graph vertices are modelled with two separate adjacency lists:
44 	 * Adjacencies of a graph node are either parent (older) commits,
45 	 * and child (younger) commits.
46 	 */
47 	int nparents;
48 	struct got_object_id_queue parent_ids;
49 	int nchildren;
50 	struct got_object_id_queue child_ids;
51 
52 	time_t commit_timestamp;
53 
54 	/* Used during graph iteration. */
55 	TAILQ_ENTRY(got_commit_graph_node) entry;
56 };
57 
58 TAILQ_HEAD(got_commit_graph_iter_list, got_commit_graph_node);
59 
60 struct got_commit_graph {
61 	/* The set of all commits we have traversed. */
62 	struct got_object_idset *node_ids;
63 
64 	/* The commit at which traversal began (youngest commit in node_ids). */
65 	struct got_commit_graph_node *head_node;
66 
67 	/*
68 	 * A set of object IDs of known parent commits which we have not yet
69 	 * traversed. Each commit ID in this set represents a branch in commit
70 	 * history: Either the first-parent branch of the head node, or another
71 	 * branch corresponding to a traversed merge commit for which we have
72 	 * not traversed a branch point commit yet.
73 	 *
74 	 * Whenever we add a commit with a matching ID to the graph, we remove
75 	 * its corresponding element from this set, and add new elements for
76 	 * each of that commit's parent commits which were not traversed yet.
77 	 *
78 	 * When API users ask us to fetch more commits, we fetch commits from
79 	 * all currently open branches. This allows API users to process
80 	 * commits in linear order even though the history contains branches.
81 	 */
82 	struct got_object_idset *open_branches;
83 
84 	/* The next commit to return when the API user asks for one. */
85 	struct got_commit_graph_node *iter_node;
86 
87 	/* The graph iteration list contains all nodes in sorted order. */
88 	struct got_commit_graph_iter_list iter_list;
89 };
90 
91 static struct got_commit_graph *
92 alloc_graph(void)
93 {
94 	struct got_commit_graph *graph;
95 
96 	graph = calloc(1, sizeof(*graph));
97 	if (graph == NULL)
98 		return NULL;
99 
100 	graph->node_ids = got_object_idset_alloc();
101 	if (graph->node_ids == NULL) {
102 		free(graph);
103 		return NULL;
104 	}
105 
106 	graph->open_branches = got_object_idset_alloc();
107 	if (graph->open_branches == NULL) {
108 		got_object_idset_free(graph->node_ids);
109 		free(graph);
110 		return NULL;
111 	}
112 
113 	TAILQ_INIT(&graph->iter_list);
114 	return graph;
115 }
116 
117 #if 0
118 static int
119 is_head_node(struct got_commit_graph_node *node)
120 {
121 	return node->nchildren == 0;
122 }
123 
124 static int
125 is_merge_point(struct got_commit_graph_node *node)
126 {
127 	return node->nparents > 1;
128 }
129 
130 int
131 is_branch_point(struct got_commit_graph_node *node)
132 {
133 	return node->nchildren > 1;
134 }
135 
136 static int
137 is_root_node(struct got_commit_graph_node *node)
138 {
139 	return node->nparents == 0;
140 }
141 #endif
142 
143 static void
144 add_node_to_iter_list(struct got_commit_graph *graph,
145     struct got_commit_graph_node *node,
146     struct got_commit_graph_node *child_node)
147 {
148 	struct got_commit_graph_node *n, *next;
149 	
150 	if (TAILQ_EMPTY(&graph->iter_list)) {
151 		TAILQ_INSERT_TAIL(&graph->iter_list, node, entry);
152 		return;
153 	}
154 
155 	/*
156 	 * If a child node is known, start iterating there.
157 	 * All parent commits *should* appear before their children unless
158 	 * commit timestamps are broken (in which case the ordering of
159 	 * commits will be broken either way).
160 	 */
161 	n = child_node ? child_node : TAILQ_FIRST(&graph->iter_list);
162 	do {
163 		if (node->commit_timestamp < n->commit_timestamp) {
164 			next = TAILQ_NEXT(n, entry);
165 			if (next == NULL) {
166 				TAILQ_INSERT_AFTER(&graph->iter_list, n,
167 				    node, entry);
168 				break;
169 			}
170 			if (node->commit_timestamp >= next->commit_timestamp) {
171 				TAILQ_INSERT_BEFORE(next, node, entry);
172 				break;
173 			}
174 		} else {
175 			TAILQ_INSERT_BEFORE(n, node, entry);
176 			break;
177 		}
178 		n = TAILQ_NEXT(n, entry);
179 	} while (n);
180 }
181 
182 static const struct got_error *
183 add_vertex(struct got_object_id_queue *ids, struct got_object_id *id)
184 {
185 	struct got_object_qid *qid;
186 
187 	qid = calloc(1, sizeof(*qid));
188 	if (qid == NULL)
189 		return got_error_from_errno();
190 
191 	qid->id = got_object_id_dup(id);
192 	if (qid->id == NULL) {
193 		const struct got_error *err = got_error_from_errno();
194 		free(qid);
195 		return err;
196 	}
197 
198 	SIMPLEQ_INSERT_TAIL(ids, qid, entry);
199 	return NULL;
200 }
201 
202 static const struct got_error *
203 add_node(struct got_commit_graph_node **new_node,
204     struct got_commit_graph *graph, struct got_object_id *commit_id,
205     struct got_commit_object *commit, struct got_commit_graph_node *child_node)
206 {
207 	const struct got_error *err = NULL;
208 	struct got_commit_graph_node *node, *existing_node;
209 	struct got_object_qid *qid;
210 
211 	*new_node = NULL;
212 
213 	node = calloc(1, sizeof(*node));
214 	if (node == NULL)
215 		return got_error_from_errno();
216 
217 	memcpy(&node->id, commit_id, sizeof(node->id));
218 	SIMPLEQ_INIT(&node->parent_ids);
219 	SIMPLEQ_INIT(&node->child_ids);
220 	SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
221 		err = add_vertex(&node->parent_ids, qid->id);
222 		if (err)
223 			return err;
224 		node->nparents++;
225 	}
226 	node->commit_timestamp = commit->committer_time; /* XXX not UTC! */
227 
228 	err = got_object_idset_add((void **)(&existing_node),
229 	    graph->node_ids, &node->id, node);
230 	if (err == NULL) {
231 		struct got_object_qid *qid;
232 
233 		add_node_to_iter_list(graph, node, child_node);
234 		err = got_object_idset_remove(graph->open_branches, commit_id);
235 		if (err && err->code != GOT_ERR_NO_OBJ)
236 			return err;
237 		SIMPLEQ_FOREACH(qid, &commit->parent_ids, entry) {
238 			if (got_object_idset_get(graph->node_ids, qid->id))
239 				continue; /* parent already traversed */
240 			err = got_object_idset_add(NULL, graph->open_branches,
241 			    qid->id, node);
242 			if (err && err->code != GOT_ERR_OBJ_EXISTS)
243 				return err;
244 		}
245 		*new_node = node;
246 	} else if (err->code == GOT_ERR_OBJ_EXISTS) {
247 		err = NULL;
248 		free(node);
249 		node = existing_node;
250 	} else {
251 		free(node);
252 		return err;
253 	}
254 
255 	if (child_node) {
256 		struct got_object_qid *cid;
257 
258 		/* Prevent linking to self. */
259 		if (got_object_id_cmp(commit_id, &child_node->id) == 0)
260 			return got_error(GOT_ERR_BAD_OBJ_ID);
261 
262 		/* Prevent double-linking to the same child. */
263 		SIMPLEQ_FOREACH(cid, &node->child_ids, entry) {
264 			if (got_object_id_cmp(cid->id, &child_node->id) == 0)
265 				return got_error(GOT_ERR_BAD_OBJ_ID);
266 		}
267 
268 		err = add_vertex(&node->child_ids, &child_node->id);
269 		if (err)
270 			return err;
271 		node->nchildren++;
272 	}
273 
274 	return err;
275 }
276 
277 const struct got_error *
278 got_commit_graph_open(struct got_commit_graph **graph,
279     struct got_object_id *commit_id, struct got_repository *repo)
280 {
281 	const struct got_error *err = NULL;
282 	struct got_commit_object *commit;
283 
284 	*graph = NULL;
285 
286 	err = got_object_open_as_commit(&commit, repo, commit_id);
287 	if (err)
288 		return err;
289 
290 	*graph = alloc_graph();
291 	if (*graph == NULL) {
292 		got_object_commit_close(commit);
293 		return got_error_from_errno();
294 	}
295 
296 	err = add_node(&(*graph)->head_node, *graph, commit_id, commit, NULL);
297 	got_object_commit_close(commit);
298 	if (err) {
299 		got_commit_graph_close(*graph);
300 		*graph = NULL;
301 		return err;
302 	}
303 	
304 	return NULL;
305 }
306 
307 struct got_commit_graph_branch {
308 	struct got_object_id parent_id;
309 	struct got_commit_graph_node *node;
310 };
311 
312 struct gather_branches_arg {
313 	struct got_commit_graph_branch *branches;
314 	int nbranches;
315 };
316 
317 static void
318 gather_branches(struct got_object_id *id, void *data, void *arg)
319 {
320 	struct gather_branches_arg *a = arg;
321 	memcpy(&a->branches[a->nbranches].parent_id, id, sizeof(*id));
322 	a->branches[a->nbranches].node = data;
323 	a->nbranches++;
324 }
325 
326 static const struct got_error *
327 fetch_commits_from_open_branches(int *ncommits, int *wanted_id_added,
328     struct got_commit_graph *graph, struct got_repository *repo,
329     struct got_object_id *wanted_id)
330 {
331 	const struct got_error *err;
332 	struct got_commit_graph_branch *branches;
333 	struct gather_branches_arg arg;
334 	int i;
335 
336 	*ncommits = 0;
337 	if (wanted_id_added)
338 		*wanted_id_added = 0;
339 
340 	arg.nbranches = got_object_idset_num_elements(graph->open_branches);
341 	if (arg.nbranches == 0)
342 		return NULL;
343 
344 	/*
345 	 * Adding nodes to the graph might change the graph's open
346 	 * branches state. Create a local copy of the current state.
347 	 */
348 	branches = calloc(arg.nbranches, sizeof(*branches));
349 	if (branches == NULL)
350 		return got_error_from_errno();
351 	arg.nbranches = 0; /* reset; gather_branches() will increment */
352 	arg.branches = branches;
353 	got_object_idset_for_each(graph->open_branches, gather_branches, &arg);
354 
355 	for (i = 0; i < arg.nbranches; i++) {
356 		struct got_object_id *commit_id;
357 		struct got_commit_graph_node *child_node, *new_node;
358 		struct got_commit_object *commit;
359 
360 		commit_id = &branches[i].parent_id;
361 		child_node = branches[i].node;
362 
363 		err = got_object_open_as_commit(&commit, repo, commit_id);
364 		if (err)
365 			break;
366 
367 		err = add_node(&new_node, graph, commit_id, commit, child_node);
368 		got_object_commit_close(commit);
369 		if (err) {
370 			if (err->code != GOT_ERR_OBJ_EXISTS)
371 				break;
372 			err = NULL;
373 		}
374 		if (new_node)
375 			(*ncommits)++;
376 		if (wanted_id && got_object_id_cmp(commit_id, wanted_id) == 0)
377 			*wanted_id_added = 1;
378 	}
379 
380 	free(branches);
381 	return err;
382 }
383 
384 const struct got_error *
385 got_commit_graph_fetch_commits(int *nfetched, struct got_commit_graph *graph,
386     int limit, struct got_repository *repo)
387 {
388 	const struct got_error *err;
389 	int ncommits;
390 
391 	*nfetched = 0;
392 
393 	while (*nfetched < limit) {
394 		err = fetch_commits_from_open_branches(&ncommits, NULL,
395 		    graph, repo, NULL);
396 		if (err)
397 			return err;
398 		if (ncommits == 0)
399 			break;
400 		*nfetched += ncommits;
401 	}
402 
403 	return NULL;
404 }
405 
406 const struct got_error *
407 got_commit_graph_fetch_commits_up_to(int *nfetched,
408     struct got_commit_graph *graph, struct got_object_id *wanted_id,
409     struct got_repository *repo)
410 {
411 	const struct got_error *err;
412 	int ncommits, wanted_id_added = 0;
413 
414 	*nfetched = 0;
415 
416 	if (got_object_idset_get(graph->node_ids, wanted_id) != NULL)
417 		return NULL;
418 
419 	while (!wanted_id_added) {
420 		err = fetch_commits_from_open_branches(&ncommits,
421 		    &wanted_id_added, graph, repo, wanted_id);
422 		if (err)
423 			return err;
424 		if (ncommits == 0)
425 			return NULL;
426 		*nfetched += ncommits;
427 	}
428 
429 	return NULL;
430 }
431 
432 static void
433 free_graph_node(struct got_object_id *id, void *data, void *arg)
434 {
435 	struct got_commit_graph_node *node = data;
436 	while (!SIMPLEQ_EMPTY(&node->child_ids)) {
437 		struct got_object_qid *child = SIMPLEQ_FIRST(&node->child_ids);
438 		SIMPLEQ_REMOVE_HEAD(&node->child_ids, entry);
439 		free(child);
440 	}
441 	free(node);
442 }
443 
444 void
445 got_commit_graph_close(struct got_commit_graph *graph)
446 {
447 	got_object_idset_free(graph->open_branches);
448 	got_object_idset_for_each(graph->node_ids, free_graph_node, NULL);
449 	got_object_idset_free(graph->node_ids);
450 	free(graph);
451 }
452 
453 const struct got_error *
454 got_commit_graph_iter_start(struct got_commit_graph *graph,
455     struct got_object_id *id)
456 {
457 	struct got_commit_graph_node *start_node;
458 
459 	start_node = got_object_idset_get(graph->node_ids, id);
460 	if (start_node == NULL)
461 		return got_error(GOT_ERR_NO_OBJ);
462 
463 	graph->iter_node = start_node;
464 	return NULL;
465 }
466 
467 const struct got_error *
468 got_commit_graph_iter_next(struct got_object_id **id,
469     struct got_commit_graph *graph)
470 {
471 	*id = NULL;
472 
473 	if (graph->iter_node == NULL) {
474 		/* We are done interating, or iteration was not started. */
475 		return got_error(GOT_ERR_ITER_COMPLETED);
476 	}
477 
478 	if (graph->iter_node ==
479 	    TAILQ_LAST(&graph->iter_list, got_commit_graph_iter_list) &&
480 	    got_object_idset_num_elements(graph->open_branches) == 0) {
481 		*id = &graph->iter_node->id;
482 		/* We are done interating. */
483 		graph->iter_node = NULL;
484 		return NULL;
485 	}
486 
487 	if (TAILQ_NEXT(graph->iter_node, entry) == NULL)
488 		return got_error(GOT_ERR_ITER_NEED_MORE);
489 
490 	*id = &graph->iter_node->id;
491 	graph->iter_node = TAILQ_NEXT(graph->iter_node, entry);
492 	return NULL;
493 }